Contamination of food and drinking water with arsenic is a serious global health issue, as arsenic exposure increases risk for cancer, cardiovascular disease, respiratory conditions, and overall mortality. Susceptibility to arsenic toxicity is partilly determined by genetic factors that influence an individual's capacity to metabolize arsenic, a process that facilitates the excretion of arsenic in urine. Identifying such genetic factors will enable classification of individuals based on toxicity risk and elucidate the biological mechanism underlying susceptibility to arsenic toxicity, informing the development of interventions that reduce toxicity. Prior research has demonstrated that there are at least two independent association signals in the 10q24.32 region (which contains the arsenic methytransferase gene; AS3MT). However, prior studies have been unable to identify the causal variants in this region due to lack of (1) complete data on all genetic variants in the region, (2) large sample sizes from multiple population groups, and (3) comprehensive functional annotation for non-coding variants. Furthermore, the potential effects of rare variants in this region have never been assessed. We propose to fill these knowledge gaps by sequencing this region in >4,500 individuals from three arsenic- exposed population groups: Bangladeshis, Native Americans, and European Americans. Within each of these groups, we will assess associations between variants in this region and arsenic methylation capacity (i.e., urinary arsenic metabolite percentages). Variations in patterns of association across ancestry groups will allow us to narrow-in on causal variants shared across populations and examine evidence for population-specific signals. Statistical evidence of causal association will be assessed using a Bayesian approach. Evidence of functionality will be assessed using annotation of non-coding variants based on prior evidence of local transcription factor binding (ChIP-Seq), DNaseI hypersensitivity, chromatin marks, and cis-gene expression. Evidence for gene-arsenic interaction will be assessed. We will determine if rare coding variants in the AS3MT gene collectively influence arsenic methylation capacity using gene-level association tests. To assess the implications of these variants for arsenic-related health outcomes, we will test associations between the 10q24.32 variants identified in aims 1-2 and risk for arsenical skin lesions (among Bangladeshi cases and controls) and squamous cell skin cancer (among European American cases and controls). Identifying potential causal variants and assessing the effects of rare variants is a logical and essential next step for elucidating the critical role of the 10q24.32 regon in arsenic metabolism and toxicity. The knowledge we are proposing to generate will enhance risk prediction, guide the development future research and prevention efforts, and clarify the biological mechanisms that underlie inter-individual differences in susceptibility to arsenic toxicity.